1. Field of the Invention
The present invention relates to an image processing device and image processing method, which correct a deformation in a photographic image of an object to be photographed resulting from the photographic position of the object, and more particularly to an image processing device and image processing method, which reduce the blurriness or the like of an image by carrying out respectively different image processings in accordance with a correction amount for correcting the deformation of a photographic image.
2. Description of the Related Art
There is a field of technology called image processing technology for adjusting the contrast and so forth of digital images taken with a digital camera.
The image processing technology encompasses various methods. For example, there is an image processing method, which is called a spatial filter process, whereby, in order to reduce the noise, and to highlight and extract the lines and edges contained in an image to be processed, picture elements (pixels) in the image are subjected to processing to change the brightness of the pixels by carrying out either a multiply-and-accumulate operation or a nonlinear operation corresponding to the brightness of the pixels in the vicinity of the pixels to be changed. Also, there is an image processing method, which makes use of a table called a LUT (Look Up Table) to record the relationship between the input values and output values of the brightness of the pixels in an image, and changes the brightness of the pixels using the LUT. Also, there is an image processing method called shading compensation, which performs compensation for the irregularities in brightness in an image resulting from the characteristics of the optical system and imaging system at photographing time so that the image has a uniform brightness. Further, there is an image processing method called a binarizing process, which changes an image having gradation to a binary image.
Further, Japanese Patent Application Publication No. 2002-247363 discloses an image processor capable of performing a natural texture removal process even when the manuscript being scanned by a scanner is a stacked manuscript or a manuscript in which there are irregularities at the texture level. Such image processor has an image inputting unit for inputting an image; a texture level detector for detecting the texture level of an inputted image; a stacked manuscript determining unit for determining whether or not an image is a stacked manuscript based on either user instructions or a texture level detected by the texture level detector; a texture level adjusting unit for adjusting the texture level corresponding to the distribution of the texture level detected by the texture level detector, when the manuscript has been determined to be a stacked manuscript; and a texture removal processor for removing the texture of an image at the texture level adjusted by the texture level adjusting unit, is provided so as to enable.
Furthermore, as an example of digital image usage, the practice of using a digital camera to take a photograph of the contents of a whiteboard being used at a meeting to retain as information has become widespread recently. In taking a photograph of the whiteboard, which is a flat object, with a digital camera, the physical relationship between the photographing surface of the digital camera and the whiteboard is ordinarily not parallel, and a photographic position is ordinarily in a position that they are not facing one another. Therefore, parallel straight lines, which form squares on the whiteboard, are photographed distortedly, i.e., not parallelly, in the photographic image, and characters, symbols and so forth written on the whiteboard are photographed in a distorted state, since and the whiteboard is distortedly photographed.
In order to correct the photographic image in which the object is distortedly photographed, Japanese Patent Application Publication No. 2005-38021 provides an image processing apparatus, an image input device, an image processing method, and a program for executing the image processing method on a computer, in which a tilt correction is conducted to another image photographed from only one view point by calculating or selecting a reference front image acquired by photographing the plane-like object and using the front image.
Further, in order to generate via a simple operation a corrected image with no distortions or deformations, from a photographic image in which an object is distortedly photographed, Japanese Patent Application Publication No. 2001-84365 provides an image tilt correction method, which has a step for determining the coordinates of two points corresponding to two end points, which are obtained by projecting the two end points of a first straight line onto a vertical line, which passes through the midpoint of the first straight line of a deformation diagram in a user-specified photographic image; a step for determining the coordinates of two points corresponding to two end points, which are obtained by projecting the two end points of a second straight line onto a vertical line, which passes through the midpoint of the second straight line of a diagram in a user-specified photographic image; a step for determining a radioscopic conversion matrix such that the two end points of a first straight line and the two end points of a second straight line are transformed to the corresponding points, which were determined in the above-mentioned steps; and a step for radioscopically converting the original image based on the obtained radioscopic transformation matrix.
However, for a photographic image in which a distorted photographic image has been photographed, when the distortion in the photographic image is corrected using the technologies described in the Japanese Patent Application Publication No. 2005-38021 and Japanese Patent Application Publication No. 2001-84365, a part of the image data in the image is enlarged as shown in FIGS. 15A through 15C, and FIG. 16. In these technologies, however, in a portion with a higher enlargement ratio, a larger amount of image data is enlarged, and therefore the image tends to become blurry.